We formulate and test a theory of posttraumatic stress disorder (PTSD) based only on mechanisms that have solid support in scientific studies of cognition, personality, and emotion. The theory combines findings from experimental and individual differences research to make predictions that often contradict existing theories of PTSD, but that agree with the existing data. It is the first theory to combine measures of the severity of traumatic events, personality, and memory processes. Moreover, it predicts quantitatively clinical outcomes by assigning numerical weights to the various measures. Two main studies are proposed, which use structural equation models. In the first, 200 adults will be tested, who vary in symptom severity. Of these, at least 50 will meet standard clinical diagnostic inclusion criteria for current PTSD and 50 will meet standard exclusion criteria for lifetime PTSD so that, in addition to the correlational analyses, direct comparisons can be made between the prototypical groups used in the PTSD literature. To provide predictor variables, the participants will complete standardized tests of mood and personality. They will recall and record a narrative version of their most traumatic or stressful events, which will be scored for coherence and event severity. They will also recall their three most positive and three most important events and imagine that three traumatic scenarios occur to them in the future. The imagined future events will probe how the participants view and react to traumatic events in general. For each of their memories and the imagined events, participants will provide measures of emotional intensity and valence, vividness of sensory information, narrative coherence, and degree of reliving of the event. They will also provide ratings of involuntary, often intrusive, memories made as they come to mind for two weeks, recording their responses on personal data assistants. This online technique was developed by our group and its use is still unique to it in PTSD populations. It provides the most accurate measure of intrusive memories, a key symptom of PTSD. The rich data set obtained allows us to disentangle competing existing hypotheses and better understand the mechanisms that lead to and maintain PTSD. The second study uses the same theory and quantitative techniques, but takes advantage of an ongoing longitudinal study in its 20th year. It includes fewer, but similar, tests to the first study and a sample of several thousand adults approximately 60 years old. It includes tests of current PTSD symptom severity, mood and personality tests, as well as health data, to do a truly prospective study of current PTSD symptoms as predicted by measures obtained longitudinally. Additional tests and experiments will be conducted to test specific supporting hypotheses. In particular, we will examine the coherence of narrative of memories of traumatic, very positive, and important events in participants who have or do not have a clinical diagnosis of PTSD.